Running a large number of files from a corpus allocates a lot more memory

[gustavo-grieco]

When the contracts to run are large and the number of sequences of transactions in the corpus is large (e.g. 500), reloading the corpus produces 2x or 3x mores memory than expected.